The Web Site for Critical Realism The WSCR Archive
[The following paper is archived here with the permission of the author, who retains the copyright.]

Notes for a Reading Group: A Short Introduction to the Critical Realist Approach to Science


1. Introduction

During the last reading group we spent a fair bit of time debating empiricism, the empirical method, and prediction. I thought I would have a go at outlining what I perceive to be a Critical Realist approach to these notions. The solution lies in a return to Bhaskar's first book A Realist Theory of Science. This is, according to many, his greatest work. I actually did not start my Critical Realist studies with it, and in many ways I know it much less well than his other books. Part of the reason is that the Bhaskarian movement is roughly speaking divided into Critical Realists and Dialectical Critical Realists. Again, broadly speaking, these groupings can be thought as clustering around the canonical moments of A Realist Theory of Science and Dialectic: The Pulse of Freedom respectively. I, of course, belong to the Dialectic camp. But there is a unity about Bhaskar's work, and I constantly run into the reality that A Realist Theory of Science is an essential starting point.

2. The Domains of the Real, the Actual, and the Empirical

In his introduction to A Realist Theory of Science, Bhaskar outlines what he terms three domains: the real, the actual, and the empirical. The real consists of underlying structures and mechanisms, and relations; events and behaviour; and experiences. The structures and mechanisms generate events in the natural world. Relations generate behaviour in the social world. The domain of the actual consists of these events and behaviour. The domain of the empirical consists of what we experience.

It is interesting to note that it was Bhaskar's distinction between the domain of the real and the domain of the actual that was hailed by his teacher Rom Harre as one of the great philosophical discoveries of the 20th century.

It is important, I think, to understand that the distinction between the three domains allows one to argue that events will take place whether they are observed or not. The tree will grow whether we observe it or not. The argument of course goes back to Bishop Berkeley (1695-1753). He reduced the domains of the real and the actual to the empirical when he said that nothing in the world existed if it was not perceived. When there were no humans around to do the perceiving, the gap was filled in, according to Berkeley, by God. This gave rise to the famous limericks by Ronald Knox:

There was a young man who said, "God
Must think it exceedingly odd
If he finds that this tree
Continues to be
When there's no-one about in the Quad."

And the reply

Dear Sir, Your astonishment's odd
I am always about in the quad.
And that's why this tree
Will continue to be,
Since observed by
Yours faithfully,
God. (cited in Lewis 54)

3. Laws and Predictions; Closed and Open Systems

In some ways, it is easy to make fun of Berkeley, but his arguments keep cropping up again and again, as does his solution of resorting to God or some transcendent to fill in the gap caused by our failure to acknowledge nature as existing independently of the human. Think here of Graeme Turner's remarks about the relationship between language and reality. Turner has claimed that "language does not describe reality, it actually constitutes it" (Turner cited in Windschuttle, "The Poverty of Media Theory" 18, Turner's emphasis). In other words, man is now god and constructs nature.

The deep confusion in Turner's position has been ruthlessly exploited by the neo-positivist Keith Windschuttle in the debate about the status of Media Theory and its relevance to the education of journalists (see Windschuttle, "The Poverty of Media Theory" and "Cultural Studies Versus Journalism"). In "Cultural Studies Versus Journalism," for instance, Windschuttle describes Turner's views as "crude and stupid" (18). The pot has a go at the kettle!

Turner's views are of course far from crude as any perusal of his account of Media Studies in Cunningham and Turner's The Media in Australia would show (Turner 203-64). Nevertheless, Turner's remark about the relationship between language and reality is something like a combination of Kantian reduction of ontology to epistemology and possibly a Rortyan postmodernism. He would, however, have done better to say that we construct knowledge through language, but that reality exists independent of our constructions of it.

Another area where Critical Realism diverges radically from the dominant Philosophy of Science approach, exemplified by Popper, is in the definition of and role of laws. For Popper and Hempel, we have explained an event when we have formulated a universal law from which the event can be deduced. The basic form of the law is: "if x then y". So prediction is built into its formulation. If we can identify the x, then we can predict that y will follow. This also gives us the essence of empiricism, namely the constant conjunction of events--if x then y.

Bhaskar's contribution here was to point out that constant conjunctions of events only occur in closed systems. Indeed, the definition Bhaskar gives of a closed system is one where the constant conjunction of events occurs (A Realist Theory of Science 69). He also argued that such closed systems do not occur naturally but are the outcome of a good deal of work on the part of a scientist.

Bhaskar further pointed out that the work necessary to produce a closed system was a transcendental proof that the world was, in fact, open but susceptible to regional closures. In The Possibility of Naturalism, he went on to argue that in the social sciences it is impossible to produce a closed system. Accordingly, the Humean formula of "if x then y" does not apply in the social sciences.

We will eventually come to this in our reading. We spent a good deal of time around the question of prediction attached to notions that TV causes violent behaviour; in other words, if a child watches violent television (x), then he will behave violently (y). But here, if I may, I will take as an example the notorious Bell shaped curve which plagues our lives every semester. This as applied in our school takes the form of if x, i.e. a class of more than 25, then y, i.e. 15% will get a 7, 25% will get a 5 etc. Lecturers who produce results that conform to this formula have their results passed without a murmur. Those who failed to conform to the formula have their results subject to scrutiny and, on occasion, scorn.

In Bhaskarian terms, when we apply the curve we are claiming that this is ontological determinism. That is, we are saying that reality is like this. Our results, if they are to reflect reality, must conform to the curve. So apparently at present within the Arts faculty, there is a campaign being mounted around the notion that there are too many "fives" in the results from the Arts faculty. The basis for saying that there are too many fives is ostensibly the curve. Our results must be wrong for they violate an ontological principle--they are in defiance of how reality determines things. However, what in truth is happening is not ontological determinism, for social reality does not conform to a curve.

The notion of a curve assumes that there is a symmetry between natural and social systems in that it ignores the factor of human agency. It says in effect that it does not matter how hard or not a particular group of students may study, the results will be the same. If they are not, then the lecturer has erred. It thus assumes implicitly that classes are closed systems where the constant conjunction of events must apply. I have stressed "implicitly" because, of course, it is impossible in QUT to get from anyone a theoretical statement justifying the curve. In the way of positivism, no theory is needed because that is how things just are. However, any class is an open system with a variety of patterns of human agency. So no single pattern of results is likely. What we have instead with our adoption of the curve is epistemological pre-determination. We determine in advance what our results will be and when they do not conform to this, and they almost never do, we moderate, i.e. fake, them.

4. Three Approaches to the P>hilosophy of Science: Empiricism, neo-Kantianism and Transcendental or Critical Realism.

We had something of a debate as to what empiricism was and the distinction between empiricism and the empirical method. Bhaskar uses empiricism in two ways. Firstly, it covers the whole positivist Humean tradition and the neo-Kantian tradition. Basically empiricism means those systems that assume a constant conjunction of events is necessary for a law.

The second use of empiricism is in classical empiricism or empirical realism as a particular philosophy of science. This identifies regularities whose constant conjunction form facts.

The rival tradition of neo-Kantianism stresses the importance of the social activity of scientists, especially the models they build of the imagined or imaginary mechanisms which would explain the regularities.

Critical Realism goes a step further than neo-Kantianism and demands that these models of imagined mechanisms be subject to empirical testing to determine whether they are real or not (A Realist Theory of Science 14-15). The neo-Kantians follow Kant in their belief that the reality of the models of the imagined mechanisms they create cannot be established.

Critical Realism differs from classical empiricism in regarding the regularities initially observed not as facts but as results of (generally) scientific experiments.

With regard to the neo-Kantian position, it is of course not always possible to test whether the model of a particular mechanism is real. But if we take the example of research into the brain, it's clear that our ability to establish the reality of particular models of mechanisms has improved radically since the advent of magnetic resonance imaging. Indeed, 80% of what we now know about how the brain works has been established within the last 5 years. Accordingly, we have to keep an open mind about what aspects of reality we will be able to identify.

So to sum up this section then, science works as follows: a) a regularity is identified (classical empiricism), b) a plausible explanation of this regularity is invented (neo-Kantianism), and c) the reality of the entities and processes that have been postulated are checked (Critical Realism) (A Realist Theory of Science 14).

5. Conclusion

Hopefully these notes will have established the fundamentals of the Critical Realist approach to science. They are of course no substitute for a reading of A Realist Theory of Science. What we are hoping to get from our reading of The Possibility of Naturalism is the uniqueness of Bhaskar's approach to social theory where he attempts to avoid the positivist tradition which would impose the so-called laws of nature (the constant conjunction of events and predicability) on social systems and the hermeneutical tradition which stresses that science has absolutely nothing to tell us about the social sciences.

It is absolutely the key to Bhaskarian thought here to see that it stresses that both the positivists and the hermeneuticists assume that science is about the establishment of constant conjunctions of events and the subsequent ability to predict. The hermeneutical tradition simply argues that science is not applicable to social systems. Both traditions, however, are working with a very impoverished view of science, a view which Bhaskar has done more than anyone to correct.

A final note on the essential radicalness of the denial of the constant conjunction of events. This allows for change that is for the new to emerge. Contrast Bhaskar here with Nietzsche and the latter's notion of the "eternal return" or the "same old, same old" and you will see how revolutionary Bhaskar truly is.


Works Cited

Bhaskar, R. A Realist Theory of Science. Brighton: Harvester Press, 1978.

-----. Dialectic: The Pulse of Freedom. London: Verso, 1993.

-----. The Possibility of Naturalism. London: Routledge, 1998.

Lewis, J. Bertrand Russell: Philosopher & Humanist. London: Lawrence & Wishart, 1968.

Turner, G. "Media Texts and Messages." The Media in Australia. Ed. Cunningham, S., & Turner. G. Sydney: Allen & Unwin, 1993. 203-64.

Windschuttle, K., "The Poverty of Media Theory." Quadrant (March 1998): 11-18.

-----. "Cultural Studies Versus Journalism." Quadrant (March 1999): 11-20.

Copyright © 1999 Gary MacLennan

The Web Site for Critical Realism The WSCR Archive